Alternate Dissociation Pathways Identified in Charge-Reduced Protein

Anal. Chem. 2010, 82, 5363–5372

Alternate Dissociation Pathways Identified in Charge-Reduced Protein Complex Ions Kevin Pagel,† Suk-Joon Hyung,‡ Brandon T. Ruotolo,‡ and Carol V. Robinson†,* University of Cambridge, Department of Chemistry, Lensfield Road, Cambridge, CB2 1EW, U.K. Tandem mass spectrometry (MS) of large protein complexes has proven to be capable of assessing the stoichiometry, connectivity, and structural details of multiprotein assemblies. While the utility of tandem MS is without question, a deeper understanding of the mechanism of protein complex dissociation will undoubtedly drive the technology into new areas of enhanced utility and information content. We present here the systematic analysis of the charge state dependent decay of the noncovalently associated complex of human transthyretin, generated by collision-induced dissociation (CID). A crown ether based charge reduction approach was applied to generate intact transthyretin tetramers with charge states ranging from 15+ to 7+. These nine charge states were subsequently analyzed by means of tandem MS and ion mobility spectrometry. Three different charge-dependent mechanistic regimes were identified: (1) common asymmetric dissociation involving ejection of unfolded monomers, (2) expulsion of folded monomers from the intact tetramer, and (3) release of C-terminal peptide fragments from the intact complex. Taken together, the results presented highlight the potential of charge state modulation as a method for directing the course of gas-phase dissociation and unfolding of protein complexes. Electrospray ionization (ESI) is one of the softest ionization methods that is currently available, and rapid technological progress, especially within the last ten years, makes ESI a core technique in the field of mass spectrometry (MS) based proteomics.1 Increasing evidence suggests that ESI-generated ions of large biomolecules can retain aspects of their solution-state conformations in the absence of bulk solvent.2 In recent years, ESI-MS has therefore evolved into a well-established tool for the characterization of noncovalently bound protein assemblies,3-8 capable of * Corresponding author. E-mail: [email protected]. † Current address: University of Oxford, Department of Chemistry, South Parks Road, Oxford, OX1 3QZ, U.K. ‡ Current address: University of Michigan, Department of Chemistry, 930 North University, Ann Arbor, MI 48109-1055. (1) Aebersold, R.; Mann, M. Nature 2003, 422, 198–207. (2) Ruotolo, B. T.; Robinson, C. V. Curr. Opin. Chem. Biol. 2006, 10, 402– 408. (3) Loo, J. A. Mass Spectrom. Rev. 1997, 16, 1–23. (4) Ashcroft, A. E. Nat. Prod. Rep. 2005, 22, 452–464. (5) Loo, J. A.; Berhane, B.; Kaddis, C. S.; Wooding, K. M.; Xie, Y.; Kaufman, S. L.; Chernushevich, I. V. J. Am. Soc. Mass Spectrom. 2005, 16, 998– 1008. (6) Wyttenbach, T.; Bowers, M. T. Annu. Rev. Phys. Chem. 2007, 58, 511– 533. 10.1021/ac101121r  2010 American Chemical Society Published on Web 05/18/2010

assessing the stoichiometry9 and topological arrangement of both soluble10 and membrane protein complexes.11 The information content of these MS experiments relies upon the retention of noncovalent interactions during ionization and transfer into the instrument. However, in many cases, to obtain information on the composition and connectivity of a complex, it is equally important to activate and dissociate the assembly in a predictable and reproducible way. Several techniques directed to this end have been described12-15 with collision-induced dissociation (CID) being the most widely employed for large protein assemblies.15-18 CID activation typically involves a large number of collisions (typically 103-105), between the ion and neutral gas atoms or molecules, that act to increase the internal energy of the ion on the hundreds of microseconds time scale.7,15 As a consequence of this “slow-heating” process, CID of protein complexes characteristically involves multiple structural rearrangement steps prior to dissociation. Current data indicate that primarily one subunit, but potentially multiple subunits, within the complex starts to unfold after collisional activation.15,19 As a result, the accessible surface area of the complex increases, and charges are transferred onto the newly exposed surface area of the unfolding subunit. This minimizes the Coulombic energy of the ion. Further activation leads to additional unfolding and charge migration until a single subunit is expelled from the protein complex leaving a “stripped” (n-1)-mer that retains 30-50% of the charges originally (7) Benesch, J. L. P.; Ruotolo, B. T.; Simmons, D. A.; Robinson, C. V. Chem. Rev. 2007, 107, 3544–3567. (8) Heck, A. J. R. Nat. Methods 2008, 5, 927–933. (9) Hernandez, H.; Robinson, C. V. Nat. Protoc. 2007, 2, 715–726. (10) Ruotolo, B. T.; Benesch, J. L. P.; Sandercock, A. M.; Hyung, S.-J.; Robinson, C. V. Nat. Protoc. 2008, 3, 1139–1152. (11) Barrera, N. P.; Isaacson, S. C.; Zhou, M.; Bavro, V. N.; Welch, A.; Schaedler, T. A.; Seeger, M. A.; Miguel, R. N.; Korkhov, V. M.; van Veen, H. W.; Venter, H.; Walmsley, A. R.; Tate, C. G.; Robinson, C. V. Nat. Methods 2009, 6, 585–587. (12) Felitsyn, N.; Kitova, E. N.; Klassen, J. S. Anal. Chem. 2001, 73, 4647– 4661. (13) Laskin, J.; Futrell, J. H. Mass Spectrom. Rev. 2005, 24, 135–167. (14) Jones, C. M.; Beardsley, R. L.; Galhena, A. S.; Dagan, S.; Cheng, G.; Wysocki, V. H. J. Am. Chem. Soc. 2006, 128, 15044–15045. (15) Benesch, J. L. P. J. Am. Soc. Mass Spectrom. 2009, 20, 341–348. (16) Light-Wahl, K. J.; Schwartz, B. L.; Smith, R. D. J. Am. Chem. Soc. 1994, 116, 5271–5278. (17) Schwartz, B. L.; Bruce, J. E.; Anderson, G. A.; Hofstadler, S. A.; Rockwood, A. L.; Smith, R. D.; Chilkoti, A.; Stayton, P. S. J. Am. Soc. Mass Spectrom. 1995, 6, 459–465. (18) Versluis, C.; van der Staaij, A.; Stokvis, E.; Heck, A. J. R.; de Craene, B. J. Am. Soc. Mass Spectrom. 2001, 12, 329–336. (19) Ruotolo, B. T.; Hyung, S.-J.; Robinson, P. M.; Giles, K.; Bateman, R. H.; Robinson, C. V. Angew. Chem. 2007, 119, 8147–8150.

Analytical Chemistry, Vol. 82, No. 12, June 15, 2010

5363

accumulated on the complex.15,20-22 The distribution of charges between the unfolded monomer and the remaining stripped oligomer is consequently highly asymmetric with respect to mass but symmetric with respect to surface area.22 For very large complexes, multiple subunit losses have been shown to proceed sequentially via the same mechanism.22 Several studies have provided evidence for a strong correlation between the overall charge state of a given protein assembly and the mechanism of its dissociation. Investigations on artificially generated peptide clusters revealed that charge state and cluster size are major determinants of their dissociation behavior.23 Further studies showed that the total charge of the precursor ion also has significant impact on the dissociation of noncovalently bound protein complexes.24,25 Protein dimers with a moderate charge state were shown to follow the common asymmetric charge partitioning pathway, whereas high-charged assemblies yielded products with a more even distribution of charge.25 Additionally, it has been shown that there is a direct correlation between the conformational flexibility of the polypeptide chain and the amount of charges carried after dissociation.24,26 Protein subunits whose conformational flexibility is limited by intramolecular disulfide bonds were shown to accommodate less charges than their nonrestricted, fully flexible analogues.24 Similar trends have been reported for heteromeric protein complexes.27 Recent investigations have also indicated that the internal energy required to dissociate a noncovalent protein assembly drops significantly with increasing charge of the parent ion, which furtherhighlights the central importance of ion charge state in the dissociation of multi-protein complexes.28 Nevertheless, many aspects of protein complex dissociation remain poorly understood, and several recent investigations provide striking evidence that some protein complexes may dissociate following very different mechanisms.27,29-31 For example, recent experiments demonstrate that tetramers of 2-keto3-deoxyarabinonate dehydratase dissociate symmetrically into dimers,29 while other proteins may dissociate without extensive unfolding, by releasing subunits carrying fewer charges,30 or even peptide fragments.30,31 To shed further light on the forces that play a critical role in dictating the mechanistic route taken for a given protein complex (20) Wanasundara, S. N.; Thachuk, M. J. Am. Soc. Mass Spectrom. 2007, 18, 2242–2253. (21) Wanasundara, S. N.; Thachuk, M. J. Phys. Chem. A 2009, 113, 3814–3821. (22) Benesch, J. L. P.; Aquilina, J. A.; Ruotolo, B. T.; Sobott, F.; Robinson, C. V. Chem. Biol. 2006, 13, 597–605. (23) Jurchen, J. C.; Garcia, D. E.; Williams, E. R. J. Am. Soc. Mass Spectrom. 2003, 14, 1373–1386. (24) Jurchen, J. C.; Garcia, D. E.; Williams, E. R. J. Am. Soc. Mass Spectrom. 2004, 15, 1408–1415. (25) Jurchen, J. C.; Williams, E. R. J. Am. Chem. Soc. 2003, 125, 2817–2826. (26) Sinelnikov, I.; Kitova, E. N.; Klassen, J. S. J. Am. Soc. Mass Spectrom. 2007, 18, 617–631. (27) Aquilina, J. A. Proteins: Struct., Funct., Bioinf. 2009, 75, 478–485. (28) Benesch, J. L. P.; Ruotolo, B. T.; Sobott, F.; Wildgoose, J.; Gilbert, A.; Bateman, R.; Robinson, C. V. Anal. Chem. 2009, 81, 1270–1274. (29) van den Heuvel, R. H. H.; van Duijn, E.; Mazon, H.; Synowsky, S. A.; Lorenzen, K.; Versluis, C.; Brouns, S. J. J.; Langridge, D.; van der Oost, J.; Hoyes, J.; Heck, A. J. R. Anal. Chem. 2006, 78, 7473–7483. (30) Dodds, E. D.; Blackwell, A. E.; Jones, C. M.; Cordes, M. H. J.; Wysocki, V. H. ASMS; American Society of Mass Spectrometry: Philadelphia, 2009. (31) Wysocki, V. H.; Jones, C. M.; Galhena, A. S.; Blackwell, A. E. J. Am. Soc. Mass Spectrom. 2008, 19, 903–913.

5364


dissociation, we analyzed systematically the charge state dependent CID behavior of tetrameric human transthyretin (TTR) by both tandem mass spectrometry and ion mobility mass spectrometry (IM-MS). We selected transthyretin (TTR) for this study since it is a very well-characterized 127-residue protein, implicated in human amyloid diseases.32-34 In its native form, TTR is stable as a 55 kDa homotetramer. Its CID induced unfolding19,35 and dissociation35-37 behavior have been studied extensively. To generate a range of charge states from this protein complex, a crown ether based charge reduction approach was applied to qualitatively and quantitatively monitor unfolding and dissociation. Our results reveal that very distinct dissociation pathways can be accessed from a single protein complex, simply by modulating the overall charge state. In a more general context, these data imply that charge reduction can be used to increase the information content of MS experiments aimed at determining macromolecular structures. MATERIALS AND METHODS Reagents, Protein Expression, and Sample Preparation. Human TTR was expressed and purified as described previously.34 The preparation was analyzed using nano-ESI-MS, and the molecular mass of a single subunit corresponded to that calculated for the unmodified sequence of amino acids in the protein. On the day of analysis, TTR stock solutions in a buffer containing 25 mM Tris phosphate (pH 8.0) and 200 mM NaCl were buffer exchanged twice into 5 mM NH4OAc (pH 7.0) using micro Biospin 6 columns (Bio-Rad, Hemel Hempstead, UK). Reagents and buffers were purchased from Sigma Aldrich (St. Louis, MO, USA). After photometric concentration determination (ε280 ) 77 600 M cm-1) on a Picodrop Microliter UV/vis spectrophotometer (GRI, Braintree, UK), samples were diluted to give a working solution of 50 µM TTR (monomer). Prior to measurements, the working solution was further diluted with buffer and/or crown ether solution to yield a final protein concentration of 2.5 µM TTR tetramer in (A) 10 mM NH4OAc (pH 7.0), (B) 10 mM triethylammonium acetate (TEAA, pH 7.0), or (C) 10 mM TEAA (pH 7.0) + 250 µM Aza-18-crown-6 (A18C6), respectively. Charge Reduction. Several strategies to reduce the charge state of ESI-generated ions of biological macromolecules have been reported within the last ten years with considerable success.38-44 These include either instrument modification or addition of considerable quantities of solution additives. The former, while an attractive method, can be difficult to implement and may cause depression of signal intensity and/or instability of the electrospray signal. Additionally, excessive concentrations of (32) Westermark, P.; Sletten, K.; Johansson, B.; Cornwell, G. G. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 2843–2845. (33) Jacobson, D. R.; Buxbaum, J. N. Adv. Hum. Genet. 1991, 20, 69–123. (34) Lashuel, H. A.; Lai, Z.; Kelly, J. W. Biochemistry 1998, 37, 17851–17864. (35) Hyung, S.-J.; Robinson, C. V.; Ruotolo, B. T. Chem. Biol. 2009, 16, 382– 390. (36) Nettleton, E. J.; Sunde, M.; Lai, Z.; Kelly, J. W.; Dobson, C. M.; Robinson, C. V. J. Mol. Biol. 1998, 281, 553–564. (37) Sobott, F.; McCammon, M. G.; Robinson, C. V. Int. J. Mass Spectrom. 2003, 230, 193–200. (38) Scalf, M.; Westphall, M. S.; Krause, J.; Kaufman, S. L.; Smith, L. M. Science 1999, 283, 194–197. (39) Ebeling, D. D.; Westphall, M. S.; Scalf, M.; Smith, L. M. Anal. Chem. 2000, 72, 5158–5161.

solution additives might affect the structural properties of the protein and/or lead to increased adduct formation. For the tandem IM-MS experiments presented here, reduction in signal intensity or perturbation of structure could not be tolerated. Therefore, a novel charge reduction approach was applied. The strategy is based on combining the charge reducing effect of gas-phase bases with the binding selectivity of crown ethers (CEs). CEs are known to form stable gas-phase host-guest complexes with peptides and protein ions by coordinating with protonated guanidine or amino groups at the side chains or the N-terminus.45-48 The noncovalent attachment of CEs has therefore evolved into a valuable method for analyzing the condensed phase side-chain accessibility in proteins by MS.49,50 However, the opposite strategy, i.e., the controlled removal of CEs, could also be beneficial. CID induced removal of noncovalently attached 18-crown-6 (18C6) in the electrospray region of the instrument can be used to lower the charge state of TTR considerably (data not shown). Even more pronounced effects can be obtained by replacing 18C6 by aza-18crown-6 (A18C6) which has a perceptibly higher gas-phase proton affinity (18C6 230 kcal/mol vs A18C6 250 kcal/mol).51,52 However, due to the selective binding of the CE, only small amounts need to be added, which reduces the chance of unintended side reactions in solution. Comparison of the unfolding behavior of 10+ TTR tetramers generated by CE charge reduction and ESI alone furthermore shows that CE charge reduction is not accompanied by structural changes (Figure S3, Supporting Information). The wide range of TTR charge states generated in this study was achieved using (i) 10 mM NH4OAc solution (15+ to 13+), (ii) 10 mM TEAA40,53 (12+ to 10+), and (iii) 10 mM TEAA, 250 µM A18C6 (this equates to 100 equiv of CE to TTR) (10+ to 7+). For A18C6 containing solutions, the cone voltage was furthermore increased from 50 to 200 V. Ion Mobility-MS Experiments. IM-MS experiments were performed on a Synapt HDMS (Waters, Manchester, UK) quadrupole-ion trap-IM-MS instrument equipped with a nanoflow electrospray (nano-ESI) source. Nano-ESI capillaries were prepared using a method described previously.9 Instrument parameters were adjusted to retain noncovalent interactions in tetrameric (40) Lemaire, D.; Marie, G.; Serani, L.; Laprevote, O. Anal. Chem. 2001, 73, 1699–1706. (41) Catalina, M. I.; Heuvel, R. H. H. v. d.; Duijn, E. v.; Heck, A. J. R. Chem.sEur. J. 2005, 11, 960–968. (42) Smith, L. M. J. Am. Soc. Mass Spectrom. 2008, 19, 629–631, and literature cited therein. (43) Bagal, D.; Zhang, H.; Schnier, P. D. Anal. Chem. 2008, 80, 2408–2418. (44) Bagal, D.; Kitova, E. N.; Liu, L.; El-Hawiet, A.; Schnier, P. D.; Klassen, J. S. Anal. Chem. 2009, 81, 7801–7806. (45) Julian, R. R.; Beauchamp, J. L. Int. J. Mass Spectrom. 2001, 210-211, 613– 623. (46) Julian, R. R.; Akin, M.; May, J. A.; Stoltz, B. M.; Beauchamp, J. L. Int. J. Mass Spectrom. 2002, 220, 87–96. (47) Julian, R. R.; Beauchamp, J. L. J. Am. Soc. Mass Spectrom. 2004, 15, 616– 624. (48) Weimann, D. P.; Winkler, H. D. F.; Falenski, J. A.; Koksch, B.; Schalley, C. A. Nat. Chem. 2009, 1, 573–577. (49) Ly, T.; Julian, R. R. J. Am. Soc. Mass Spectrom. 2006, 17, 1209–1215. (50) Liu, Z.; Cheng, S.; Gallie, D. R.; Julian, R. R. Anal. Chem. 2008, 80, 3846– 3852. (51) Sharma, R. B.; Blades, A. T.; Kebarle, P. J. Am. Chem. Soc. 1984, 106, 510–516. (52) Julian, R. R.; Beauchamp, J. L. J. Am. Soc. Mass Spectrom. 2002, 13, 493– 498. (53) Verkerk, U. H.; Peschke, M.; Kebarle, P. J. Mass Spectrom. 2003, 38, 618– 631.

TTR. Typical settings were: capillary voltage, 1.0-1.5 kV; sample cone, 50 V for NH4OAc and TEAA samples, 200 V for TEAA + A18C6 samples; cone gas, off; trap and transfer collision voltage, 10 V; ion transfer stage pressure, 4.00 mbar; trap pressure, 4.82 × 10-2 mbar; ion mobility gas, N2; ion mobility cell pressure, 4.31 × 10-1 mbar; time-of-flight analyzer pressure, 1.20 × 10-6 mbar. All mass spectra were calibrated externally using an aqueous solution of cesium iodide (100 mg mL-1) and were processed with the Masslynx software (version 4.1, Waters, Manchester, UK). Spectra are shown with minimal smoothing and without background subtraction. IM-MS spectra were converted into collision cross sections (CCS)10,54,55 using an external calibration method described previously.10 Depending on the size of the analyte, different ions of known CCS were used. TTR monomers: denatured Cytochrome C and Myoglobin. TTR Trimers and tetramers: Avidin tetramer, charge states 15+ to 17+; SAP pentamer, charge states 22+ to 25+. Additionally, absolute CCS of TTR tetramers were measured on an in-house modified, linear field drift tube instrument56 and were used for internal CCS calibration. Instrument and solution conditions of these measurements were comparable to those used for the data presented here. Protein Complex Unfolding and Dissociation. Protein complex dissociation and unfolding experiments were performed by using tandem IM-MS. Individual charge states of TTR were selected in the quadrupole mass analyzer followed by acceleration in the trap region of the instrument. At increasing acceleration voltage, these mass-selected ions encounter neutral gas molecules with greater kinetic energy which leads to unfolding and/or dissociation of the parent ion. The resulting product ions are separated subsequently based on their conformation in the ion mobility cell followed by timeof-flight mass analysis.57 All measurements were repeated at least twice. The relative errors for CCS were in the range of 2-3.5%. To rule out influences of electric field on the ionmobility separation, the measurements were repeated at different wave velocities. This resulted in different drift times but, after calibration, yielded comparable CCS values. The dissociation coordinates shown represent the relative amount of intact TTR tetramer and correspond to the intensity of the tetramer divided by the intensity of all ions. For some ions, an initial dissociation coordinate below 1.0 (i.e., below 100% intact tetramer) has been obtained. This is a result of charge stripping events in the instrument.19 To consider the charge-stripped tetramer as background, all dissociation profiles were normalized regarding their intensity to cover the full range from 0 to 1. The midpoint of the dissociation energy was determined from fitting three-parameter sigmoidal curves to the normalized data sets. IM unfolding coordinates correspond to the integrated intensity of the most compact tetramer relative to the integral of the total tetramer. The midpoint of the unfolding energy was determined (54) Jarrold, M. F. Annu. Rev. Phys. Chem. 2000; 51, 179-207. (55) Hoaglund-Hyzer, C. S.; Counterman, A. E.; Clemmer, D. E. Chem. Rev. 1999, 99, 3037–3080. (56) Bush, M. F.; Giles, K.; Hoyes, J.; Robinson, C. V.; Ruotolo, B. T., private communication. (57) Pringle, S. D.; Giles, K.; Wildgoose, J. L.; Williams, J. P.; Slade, S. E.; Thalassinos, K.; Bateman, R. H.; Bowers, M. T.; Scrivens, J. H. Int. J. Mass Spectrom. 2007, 261, 1–12.


5365

similarly to the half-point dissociation energies by sigmoidal fitting of each unfolding data set. The energy applied to activate the tetramer was converted into comparable laboratory frame energy (Elab) by multiplying the acceleration voltage with the charge of the corresponding ion.13 The mass spectra of TTR in NH4OAc and TEAA buffer (charge state 15+ to 10+) were acquired at a cone voltage of 50 V, and measurements of TTR with A18C6 (charge states 10+ to 8+) required a cone voltage of 200 V. Higher cone voltages usually lead to in-source activation of TTR tetramers. As a consequence, TTR tetramer unfolds and dissociates at lower energy in the tandem MS experiment. Data sets obtained at different cone voltages can therefore only be compared quantitatively if their absolute energy axis is normalized. For the data presented, this was performed by comparing the dissociation coordinates of 10+ tetramer ions, which were generated from either 10 mM TEAA at 50 V cone or 10 mM TEAA + 250 µM A18C6 at 200 V cone. Both dissociation curves were fit to a sigmoid, and the midpoint transition values were determined. The difference between these values (118 eV) corresponds to the energy offset of the dissociation curves. To correct the energy axis of the data set obtained at a cone voltage of 200 V (charge states 10+ to 8+), this value was subtracted from each data point of the dissociation profile. RESULTS Charge Reduction. To explore a wide range of TTR charge states, it was necessary to reduce the charge in a selective and controllable manner. This has been achieved using a combination of charge-reducing buffer40,53 and a novel charge reduction approach. This latter method is based on the collision-induced removal of noncovalently attached aza-18-crown-6 from charged side chains of the protein. Further details can be found in the Materials and Methods Section. Nano-ES MS analysis of low concentrations of TTR from 10 mM ammonium acetate (AA) buffered solution under conditions chosen to maintain noncovalent interactions yields intact tetramers (Q) with charges ranging from 15+ to 13+ (Figure 1A). A moderate extent of charge reduction from 12+ to 10+ was achieved by replacing AA with triethylammonium acetate (TEAA) buffer of equal concentration (Figure 1B). The charge-reducing nature of alkylated ammonium ions has been studied extensively in recent years and is mainly based on its higher gas-phase basicity relative to AA.52 For transthyretin, ions with charges less than 10+ have been generated using the CE-based charge reduction strategy (Figure 1C). The ions shown in Figure 1A-C will form the precursor ion populations that we subject to CID in subsequent experiments. Tetramer Dissociation. To gain a detailed picture on the charge state dependence of the CID induced unfolding and dissociation of TTR tetramers, a series of tandem MS experiments was performed. Our procedure included the selection of a precursor ion in the quadrupole mass analyzer and subsequent ramping of the acceleration voltages that drive the ions into the ion trap region of the instrument. Figure 1D-F shows tandem MS spectra of 15+, 11+, and 9+ tetramers, respectively. Ions with a charge state between 15+ and 12+ followed the conventional asymmetric charge partitioning pathway by dissociating into 5366


highly charged monomers and low-charge trimers. Additionally, limited dimer dissociation was observed for 15+ ions with product ions splitting the available charge on precursor ions approximately symmetrically (Figure 1D). For the 11+ to 9+ charge state ions, the distribution of charges on the product ions was also found to be asymmetric with respect to mass, generating product ions corresponding to both monomers and trimers (Figure 1E). Interestingly, the absolute number of charges transferred to the leaving subunit decreases significantly with the lowered charge of the parent ion. Specifically, the charge state of the most abundant monomer ion decreases, from 8+ to 5+. Additionally, only marginal changes in the charge state are observed for the trimer, which shows a reduction from 7+ to 6+ for the most abundant peak. This not only indicates a change in the dissociation mechanism of low-charged TTR tetramers but also provides circumstantial evidence for the ejection of a relatively compact monomer during the dissociation process.58,59 This supposition is supported by IM measurements on the monomer ions dissociated from TTR tetramers for this charge state range (see below). Surprisingly, further changes in the dissociation mechanism are observed for ions having fewer than 9+ charges. For these ions, primarily peptide fragments are obtained as CID products (Figure 1F). Interestingly, this peptide fragmentation is not necessarily accompanied by monomer ejection from the intact TTR tetramer. Only low intensity monomer and trimer ions are observed at high collision voltages (>220 V) for the 9+ charge state (Figure 1F). Instead, mainly peptide fragments and truncated, charge-stripped TTR tetramers lacking the mass of the peptide fragments are observed. This surprising result implies that, for +9 ions of the TTR tetramer, covalent bonds within the amino acid sequence are broken at lower acceleration voltages than are needed to dissociate the noncovalently associated complex itself. Further analysis of the products reveals that most of the peptide fragments are generated by peptide bond cleavage at the Cterminal end of the sequence (Table S1, Supporting Information). The peptide product dissociation channel was even more pronounced for the 8+ tetramer, where we observed no detectable dissociation products corresponding to monomer ejection from the intact complex (Figure S1, Supporting Information). Ions with a charge of 7+ or lower produced neither peptide fragments nor ejected monomers. This indicates that the TTR tetramer ions having 7 or less charges are stable under the maximum voltages accessible on the instrumentation used here (240 V, i.e., 1680 eV for 7+ tetramers). To assess and compare the fragmentation channels described above in terms of their appearance energy, the relative intensities of the tetramer ions were plotted against the laboratory frame energies (Elab, Figure 2). The data reveal a general trend: the energy that is required to dissociate a noncovalent TTR tetramer increases with decreasing charge state. This is not surprising since highly charged ions exhibit more unfavorable Coulomb repulsions than their lowly charged analogues and consequently dissociate preferentially.28 There are however measurable differences between the dissociation profiles of each individual charge state. Tetramers with a charge of 15+ and 14+ dissociate at almost (58) Kaltashov, I. A.; Mohimen, A. Anal. Chem. 2005, 77, 5370–5379. (59) Shelimov, K. B.; Clemmer, D. E.; Hudgins, R. R.; Jarrold, M. F. J. Am. Chem. Soc. 1997, 119, 2240–2248.

Figure 1. Dissociation products of different TTR charge states. (A-C) Nano-ESI-MS spectra of 2.5 µM TTR sprayed from 10 mM NH4OAc (A), 10 mM TEAA (B), and 10 mM TEAA + 250 µM aza-18-crown-6 (C). Depending on the buffer conditions, TTR tetramer ions with a charge state of 15+ to 7+ can be generated. (D-F) Tandem IM-MS spectra of TTR charge states 15+ (D), 11+ (E), and 9+ (F) at accelerating voltages of 70, 160, and 220 V, respectively. Tetramer ions with a charge of 15+ to 12+ dissociate into highly charged monomers and charge-stripped trimers (D). 15+ ions additionally show a small fraction of symmetric dissociation into dimers. In contrast, low-charged monomers are ejected from tetramer ions with an intermediate charge state (E). Predominantly C-terminal peptide fragments are released from tetramer ions with a charge below 10+ (F). This is accompanied by conservation of the noncovalent interactions within the complex.

identical Elab (Figure 2, black and red), while a steady increase in the offset between neighboring dissociation traces is found for ions from 14+ to 12+ (Figure 2, red, green, and blue). This indicates that the impact of Coulomb repulsions on the dissociation decreases with decreasing charge state. In other words, TTR tetramers that carry relatively few charges require disproportionally more energy for their dissociation than their higher charged counterparts. However, ions that follow one of the previously described alternative pathways (11+ to 8+, Figure 2, purple, brown, green, and turquoise) exhibit a different behavior, and a near-constant difference in energy between the individual dissociation profiles is observed. Collision-Induced Unfolding of Tetrameric Transthyretin. Conformational analysis of TTR tetramer charge states (+8 to +15), and their products, was performed by ion mobility (IM).59,60 (60) von Helden, G.; Wyttenbach, T.; Bowers, M. T. Science 1995, 267, 1483– 1485.

IM separation of intact TTR 15+ tetramers at low activation energy yields a relatively narrow distribution of compact conformers with collision-cross-section (CCS) values of around 3100 Å2 (Figure 3A, lower panel).19 Upon collisional activation, the complex unfolds in a stepwise fashion leading to several partially folded intermediates, as previously observed (Figure 3A, upper panels).19,35 Contour plots of arrival time distributions across a range of activation energies for TTR charge states 15+, 13+, 11+, and 9+ are shown in Figure 3B. The collision energy at which 50% of the tetrameric complex is dissociated, as determined by tandem MS, is indicated (blue line). Contour plots for other charge states exhibit the same trend (see Supporting Information, Figure S2). Comparison of these different charge states of the TTR tetramer shows that for the lower charge states the size transitions observed change significantly. Tetramers with a charge of 15+ follow the unfolding process described previously, populating at least five intermediate states, Analytical Chemistry, Vol. 82, No. 12, June 15, 2010

5367

Figure 2. Dissociation profiles of TTR tetramers with charge states 15+ to 8+. The dissociation coordinate corresponds to the relative amount of intact tetramer. Accelerating voltages into the collision cell were converted into lab frame energies Elab and raw data fitted to a sigmoide.

each having collision-cross-section values that compare well to previous data (Figure 3B).19,35 Slightly lower charged ions (13+) unfold similarly to the 15+ with several discrete intermediates. A careful comparison of these intermediates, however, reveals a smaller increase in the CCS for the 13+ compared with the 15+. We interpret this as a lesser degree of unfolding for the 13+ charge state than for the 15+. Specifically, the most extended conformations of 15+ and 13+ tetramers have CCS of approximately 4500 and 4000 Å2, respectively. For the 11+ ions at maximum collision energy, only one slightly extended conformation with CCS of approx 3500 Å2 was observed. No discernible intermediate states were populated en route to this conformational family. Interestingly, for transthyretin tetramers with fewer than 11+ charges, no detectable increase in size was observed by ion mobility, even at the maximum activation voltages accessible (Figures S2 and S3, Supporting Information). Tetramers in the 9+ state appear to exist exclusively in a compact conformation (with a CCS of approximately 3050 Å2) under all collisional activation conditions accessible on our instrumentation (Max Elab ) 1980 V for 9+ ions, see Figure 3B). Nevertheless, these tetrameric ions can be fully dissociated to individual subunits at maximum collision energy. It can be concluded, therefore, that TTR tetramers with a charge lower than 11+ disassemble in the gas phase without the need for unfolding prior to dissociation. These data clearly indicate that the previously proposed connection between unfolding and dissociation of protein complexes does not necessarily exist for all precursor ion charge states. To enable a semiquantitative comparison between the common dissociation channel and the folded monomer ejection channel, unfolding profiles similar to those shown in Figure 2 have been extracted for charge states 15+ to 11+ (Figure 3C). It is apparent that the differences in the unfolding transition observed as a function of charge state, on the laboratory-frame energy axis, are significantly reduced for unfolding compared to dissociation (Figure 2). In other words, as the charge state on the precursor ions decreases, only small additional amounts of energy are required to bring about unfolding. This is in direct contrast to dissociation, where 5368


significantly larger amounts of energy are needed as the precursor ion charge state is decreased. Conformation of the Dissociation Products. To shed further light on the relationship between unfolding and dissociation, the conformation of product ions was characterized by analyzing the charge state partitioning between products (Figure 4A). The partitioning of charges between monomer and trimer product ions is plotted against the charge state of the tetrameric precursor ion. The average number of charges carried by both monomer and trimer products is indicated. Charge states 15+ to 12+ exhibit the expected asymmetric charge partitioning behavior with an approximately equal distribution of charges between the monomer and trimer. With decreasing parent ion charge, however, progressively less charge is transferred to the leaving subunit. For example, approximately one-third of the overall charge is transferred to monomers that originate from 9+ tetramers. In addition, it can be clearly observed that a decrease in the parent ion charge state leads to a significant decrease in the average charge of the monomer that is expelled, from 8.3 to 3.5. By contrast, only a subtle drop from 6.9 to 5.8 can be observed for the corresponding trimers. These observations therefore suggest that monomers, which originate from lowly charged TTR tetramers, possess a considerably lower surface area (i.e., a more compact conformation) than monomers which are expelled from more highly charged parent ions.58 To investigate the conformation of the trimeric and monomeric dissociation products, we employed IM analysis (Figure 4B). Results show that the CCS values of 3+ to 5+ charged monomers are significantly lower than those for their 7+ to 9+ charged analogues. The CCS value calculated using the projection approximation (PA) approach within MOBCAL61 for the X-ray structure of TTR (pdb 1F41) is approximately 20% higher than the data obtained experimentally for the 3+ to 5+ charged monomers (calcd 1140 Å2 vs exptl 1412 ± 49 Å2). Given the fact that PA calculations inherently underestimate the CCS of molecules above 200 atoms62 by approx 15%63 and considering that the structure of TTR (pdb 1F41) is lacking approximately 10% of its sequence, these data are in good agreement with CCS values expected for a compact, native-like TTR monomer. The larger CCS measured for the 7+ to 9+ monomers agree well with unfolded conformers of the transthyretin monomer (generated by molecular simulations, for details see Supporting Information) for which the CCS increases up to 100% (Table S1, Supporting Information). This indicates that monomers ejected from lowly charged TTR adopt a more compact structure compared to those originating from highly charged transthyretin tetramers. By contrast, the CCS values of the trimer are virtually unaffected by the charge state of the parent ion (2450 ± 35 Å2). This is indicative of little or no change in the conformation of the trimer dissociation product which in turn suggests limited unfolding. These findings are in good agreement with previous investigations which indicated that the charge state of the parent tetramer correlates with the extent of unfolding of the leaving (61) Mesleh, M. F.; Hunter, J. M.; Shvartsburg, A. A.; Schatz, G. C.; Jarrold, M. F. J. Phys. Chem. 1996, 100, 16082–16086. (62) Shvartsburg, A. A.; Hudgins, R. R.; Dugourd, P.; Jarrold, M. F. J. Phys. Chem. A 1997, 101, 1684–1688. (63) Scarff, C. A.; Thalassinos, K.; Hilton, G. R.; Scrivens, J. H. Rapid Commun. Mass Spectrom. 2008, 22, 3297–3304.

Figure 3. Unfolding of different TTR charge states. (A) Collisional-cross-section distributions for the 15+ charge state of TTR. Ions have been activated at trap acceleration voltages of 300, 600, and 900 eV, respectively. At low activation conditions, a single compact conformation is detected, while a heterogeneous distribution of rather extended conformers is observed at intermediate and high activation. In total, five different conformers, which have previously been denoted A-E are observed.35 (B) CCS contour plots showing the collision energy dependent unfolding of TTR charge states 15+, 13+, 11+, and 9+ (z-axis intensities in linear scale). The extent of unfolding decreases with decreasing charge. Numbers and lines depicted in blue represent the voltage at which 50% of the tetramer is dissociated. (C) Dissociation profiles of TTR charge states 15+ to 11+. The collision energy has been converted into lab frame energy Elab. Ions with charge states below 11+ do not unfold.

monomer, while the conformation of the remaining trimer remains essentially unaffected.25,26,28 Thus, the extent of charge asymmetry in the product ions observed depends upon the number of charges available for migration on the precursor ion with higher-charged precursor ions leading to more extended populations of monomer product ions. DISCUSSION Careful manipulation of the charge state of the transthyretin tetramer has enabled an in-depth study of the dissociation pathways of this protein complex and their interdependence on the initial charge state of the complex. This allowed us to follow three distinct dissociation pathways. A semiquantitative comparison between the unfolding and dissociation energetics of individual charge states is summarized (Figure 5A). The midpoint of dissociation and unfolding energies, extracted from the dissociation and unfolding traces (Figure 2 and Figure 3C), is highlighted, and the three different pathways are shaded (Figure 5A). Charge states 15+ to 12+ (pink) follow the conventional asymmetric charge partitioning pathway, which involves the gradual unfolding of one subunit and subsequent release of a highly charged, unfolded monomer carrying approximately 50% of the overall charge (Figure 5B). This is consistent with the vast majority of dissociation data reported

for protein complexes where activation of ions is achieved by slow heating. This pathway has been examined mechanistically by various experimental12,15-18,25 and theoretical studies.20,21,26 Highly charged ions are Coulombically frustrated; that is, they contain a large amount of internal tension due to the large number of charges on their surfaces and the concomitant electrostatic repulsions between those charges. Accordingly, the dissociation of these protein complex ions is driven primarily by Coulombic repulsions. Other inter- and intramolecular forces become virtually irrelevant.20 Prior to activation, the total charges on the protein complex are shared evenly between subunits of a homo-oligomeric protein. For statistical reasons, however, one subunit may carry marginally more charge on average and is therefore slightly more destabilized by intramolecular Coulombic repulsion.20 In the ground state, this has no impact on the integrity of the complex, but upon collisional activation this changes dramatically. Our data indicate that proteins carrying sufficient charge (i.e. >11+ for TTR) undergo Coulombically driven unfolding events. The extent of unfolding observed in our experiments is directly correlated with the charge of the activated precursor ion population. In accordance with previous observations, it is likely that the extent of Coulombic unfolding we observe here is Analytical Chemistry, Vol. 82, No. 12, June 15, 2010

5369

Figure 4. Conformation of CID products. (A) Relative distribution of charges on the CID products. Numbers close to the data points represent the average charge state of monomers (purple) and trimers (blue). During dissociation of highly charged tetramers, most of the protons are transferred to the ejected monomers, while only a few charges are transferred to monomers that result from low-charged tetramers. (B) Charge state dependence of CID product collision cross sections (CCS). Monomers with high charge states adopt a rather extended conformation, while compact species are found for lowcharged monomer and trimer ions.

limited only by the available charge on the precursor. Consequently, this charge migration drives unfolding until the newly exposed surface area of the monomer equates roughly to that of the folded complex that remains.22 Despite the general similarity in the dissociation pathways of the 15+ to 12+ ions, there are considerable differences in their unfolding and dissociation energetics. For charge states 15+ and 14+, there is only a small difference between the midpoint of the unfolding and dissociation values, while the difference between unfolding and dissociation energies increases for the 13+ and 12+ ions (Figure 5A). Thus, while the unfolding barrier remains essentially unaffected by the precursor ion charge, the dissociation barrier increases almost exponentially (Figure 5A). In addition, this is accompanied by a lesser extent of unfolding for moderately charged (13+ and 12+) tetramers (Figure 3B and Supporting Information Figure S1). Considering that highly charged tetramers are unfolded prior to dissociation, this is a surprising result since a comparable increase in the activation energy barrier of both processes might be expected a priori.10 Instead, our data indicate that with decreasing charge state the dissociation of the complex is no longer governed exclusively by Coulombic repulsion but rather that other intra- and intermolecular interactions become 5370


Figure 5. Distinct protein complex dissociation pathways observed for TTR charge states 15+ to 8+. (A) Charge state dependence of midpoint unfolding and dissociation values. Areas colored in purple, blue, and green indicate the dissociation pathway observed for the particular parent ion charge. Charge states 15+ to 12+ expel unfolded monomers via the common asymmetric chargepartitioning pathway (B). TTR tetramers with a charge of 11+ to 9+ dissociate into folded monomers and trimers (C). Ions below 9+ mainly release C-terminal peptide fragments without dissociation of the noncovalent complex (D).

important contributors to the protein complex dissociation energy thresholds observed here. Tetramers with a charge below 12+ dissociate into monomers and trimers but with progressively less unfolding taking place prior to dissociation as their overall charge decreases (Figure 5A and C). Only marginal changes in collision-cross-section are observed at any activation energy at which 11+ ions dissociate. Tetramers with a charge below 11+, on the other hand, do not show any evidence of unfolding under the activating conditions accessible on our instrumentation. Moreover, they retain their most compact configuration throughout the dissociation process. Additionally, there are explicit changes regarding charge partitioning and product conformation. The extent of charge migration to the leaving monomer decreases from approximately 55% to 38% for

the 12+ and 9+ ions, respectively. Consequently, the dissociation pathway can no longer be considered as strictly asymmetric with respect to the mass of the complex. Given also that the CCS values are close to the theoretical values of fully folded TTR monomers, this provides compelling evidence for a native-like structure of the departing monomer following gas-phase dissociation. It can therefore be concluded that TTR tetramers with a charge between 11+ and 9+ dissociate without considerable structural changes. If we consider this in terms of energetics, our results are consistent with an alternative pathway for which the dissociation barrier is perceptibly lower than the unfolding energy of the monomer. Interestingly, such a considerably higher unfolding barrier has been proposed theoretically for low-charged ions,21 but to date there has been little experimental evidence to support this proposal.29,30 For very low charged 9+ and 8+ tetramers, almost no dissociation into the noncovalently associated building blocks was observed. Instead, several characteristic C-terminal peptide fragments (mainly y10, y27, and y28) were detected as the major products in the tandem IM-MS experiments. It is important to emphasize that this peptide fragmentation does not result from further dissociation of already expelled monomers. Such effects have been reported previously from ions carrying very high numbers of charges28 but can be excluded here since several peaks corresponding to truncated TTR tetramers of lower charge have been observed. Interestingly, a very similar covalent fragmentation behavior has been reported recently for noncovalent SAP pentamers,31 where it was shown that conventional dissociation of SAP pentamers competes at certain conditions to a considerable extent with an alternative pathway that mainly yields characteristic peptide fragments and truncated pentamers. It is worth mentioning that the peptide fragments obtained for SAP are, similar to those obtained for charge reduced TTR, exclusively generated by backbone cleavage at the C-terminal end of the protein. The authors explain this unusual behavior as a result of the enormous accumulation of charge on the leaving subunit. The number of transferred protons might at certain conditions exceed the number of available basic residues and, therefore, lead to fragmentation.31 The results presented here suggest that this scenario is rather unlikely. For low charged TTR tetramers (11+ to 9+), there are no or only a few protons transferred to the leaving subunit (zav(monomer) ) 3.5 for 9+ tetramers). This effect will be even more pronounced for lower charge states. Additionally, the resulting peptides are only singly or doubly charged, even though there are numerous protonation sites. Currently it is unclear why peptide fragmentation occurs exclusively at the C-terminal end of the sequence, and a correlation between the internal conformation of the monomer and the cleavage sites remains elusive. A reasonable strategy to shed further light on this unresolved problem is to analyze the covalent fragmentation behavior of different TTR mutants, which are known to form tetrameric complexes with comparable stability but different internal organization.35 TTR tetramer ions with a charge below 8+ were found to be extremely stable and remained unperturbed under almost all

activation conditions accessible within the instrument. This observation may have particular impact in the proposed application of X-ray free electron laser (XFEL) imaging techniques for determining diffraction patterns of ESI-generated biomolecules in the gas phase64 as the stability of the ions generated is a crucial requirement for the success of these experiments. Additionally, there have been several recent wavelength selective IRMPD experiments aimed at obtaining structural information for gasphase peptides,65 proteins,66 and protein complexes.67 As in the X-ray diffraction experiments, the stability and internal energy of the ions observed spectroscopically will likely determine the fidelity of the results obtained. In a more general context, the experiments presented highlight the potential of charge reduction as a tool to determine experimentally the CCS of protein subunits. Knowing the CCS of individual building blocks within a protein complex is a crucial requirement for modeling the topology of an assembly using IM-MS.68 If high-resolution structural information is available, this can be achieved by calculating theoretical CCS for the subunits.61,69 However, generating models of protein assemblies is challenging when high-resolution structural information is not available.68 In these cases, the CCS of the individual subunit or subcomplex has to be determined experimentally via IM. Such measurements require a controlled disruption and at least partial dissociation of the assembly in solution.68 However, it is often difficult to identify solution conditions which facilitate dissociation of the assembly without propagating unfolding. The controlled gas-phase dissociation of charge reduced protein complex ions therefore represents an attractive alternative to measure the CCS of protein complexes and their individual subunits without unfolding or solution perturbation. CONCLUSIONS We show here that the collision-induced dissociation of gas-phase protein complexes is a highly charge state dependent process. The results presented here reveal that a controlled reduction of the charge state can significantly increase the gasphase stability of protein complexes to both unfolding and dissociative transitions. These results therefore have important consequences for hybrid structural biology methods:70 primarily, since reduction of the charge appears to protect the native state of the protein, increasing its stability for further structural interrogation. Additionally, the ability to dissociate compact monomeric subunits in the gas phase has important consequences for modeling protein complexes. In the future, charge reduction is likely to become a vital tool for the many (64) Neutze, R.; Wouts, R.; van der Spoel, D.; Weckert, E.; Hajdu, J. Nature 2000, 406, 752–757. (65) Polfer, N. C.; Oomens, J.; Suhai, S.; Paizs, B. J. Am. Chem. Soc. 2007, 129, 5887–5897. (66) Oomens, J.; Polfer, N.; Moore, D. T.; Meer, L. v. d.; Marshall, A. G.; Eyler, J. R.; Meijer, G.; von Helden, G. Phys. Chem. Chem. Phys. 2005, 7, 1345– 1348. (67) Pagel, K.; Kupser, P.; Bierau, F.; Polfer, N. C.; Steill, J. D.; Oomens, J.; Meijer, G.; Koksch, B.; von Helden, G. Int. J. Mass Spectrom. 2009, 283, 161–168. (68) Pukala, T. L.; Ruotolo, B. T.; Zhou, M.; Politis, A.; Stefanescu, R.; Leary, J. A.; Robinson, C. V. Structure 2009, 17, 1235–1243. (69) Shvartsburg, A. A.; Jarrold, M. F. Chem. Phys. Lett. 1996, 261, 86–91. (70) Robinson, C. V.; Sali, A.; Baumeister, W. Nature 2007, 450, 973–982.


5371

structural biology applications in which maintaining the native state of a protein complex and its individual subunits is imperative.

ACKNOWLEDGMENT The authors gratefully acknowledge M.G. McCammon for her assistance in protein expression and purification. Justin L.P. Benesch, Matthew F. Bush, and Joanna Freeke are acknowledged for critical reading of the manuscript and fruitful discussions. K.P. thanks the German Academy of Sciences Leopoldina for financial support. C.V.R. is a Professor of the Royal Society. B.T.R. was

5372


supported by a Waters Research Fellowship over the duration of this project. SUPPORTING INFORMATION AVAILABLE Detailed assignment of CID peptide fragments, tandem MS spectra of 8+ TTR tetramers, unfolding contour plots for charge states 15+ to 8+, and calculated CCS values of unfolded monomers generated by molecular modeling. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review April 29, 2010. Accepted May 3, 2010. AC101121R

Alternate Dissociation Pathways Identified in Charge-Reduced Protein

Recommend Documents